ANTICANCER RESEARCH 35: 4183-4190 (2015)

Reliability of Endocervical Curettage in the Diagnosis of High-grade Cervical Neoplasia and Cervical Cancer in Selected Patients

CHERIF AKLADIOS1, LISE LECOINTRE1, EMMANUELLE BAULON1, VERONIQUE THOMA1, GERLINDE AVEROUS2, MURIEL FENDER3, FRANÇOIS LEFEBVRE4 and JEAN-JACQUES BALDAUF1

Departments of 1Obstetrics and Gynecology, 2Pathology, University Hospital of Hautepierre, Strasbourg, France; 3EVE Association, Graffenstaden, France; 4Methodology and Biostatistics Sector, Public Health Department, Strasbourg University Hospitals, Strasbourg, France

Abstract. Aim: To analyze the reliability of endocervical involvement, represents a real difficulty for the colposcopist curettage (ECC) in the diagnosis of high-grade cervical trying to achieve complete visualization, as well as accurate intraepithelial neoplasia and cervical cancer, and to identify sampling by directed biopsy. In such cases, for many risk factors associated with diagnostic underestimation. colposcopists endocervical curettage (ECC) increases the Materials and Methods: A retrospective study was carried-out reliability of the examination and reduces the risk of an involving 445 patients who underwent ECC for: endocervical occult cancer (2-4). Many authors refer to situations where lesion incompletely visible on colposcopy or inaccessible to only EEC helped establish the diagnosis of cancer, either by biopsy; atypical glandular cells on smear, or discrepancy immediate analysis of the curetted sample or secondarily to between colposcopic impression and cytological abnormalities. the analysis of the cone biopsy performed because EEC Results: Comparison between ECC and final diagnosis showed revealed high-grade CIN (1-3, 5-7). Conversely, some a perfect match in 362 patients (81.3%). For 189 patients with authors blame the pain associated with ECC, as well as the pre-cancerous or cancerous endocervical lesions, the sensitivity, frequent difficulties of the pathological interpretation of specificity, and positive and negative predictive values were small fragments without stroma (8-10). Others point-out that 87.3%, 96.9%, 95.4% and 91.9%, respectively. No clinical, ECC is especially contributory in patients who are candidates cytological or colposcopic characteristics were associated with for treatment by excision (11, 12). significantly increased risk of diagnostic underestimation with The aim of our study was to analyze the reliability of ECC. Conclusion: ECC is a very reliable tool for reducing the ECC in the diagnosis of high-grade CIN and cervical number of unnecessary treatments, without increasing the risk cancer, and to identify the risk factors associated with of allowing some lesions to evolve into cancer. diagnostic underestimation in order to provide a safe approach to all patients. In the case of abnormal Papanicolau smear, the two primary objectives of colposcopy are: to identify high-grade cervical Materials and Methods intraepithelial neoplasia (CIN) in order to prevent its progression into an invasive lesion, and to identify This retrospective study included 445 patients who underwent an carcinomas to allow their early treatment. The frequency of ECC in the course of colposcopy and for whom the final endocervical involvement increases with the severity of the histological diagnosis of the endocervix was validated either by the cervical lesion (1, 2). At the same time, the endocervical analysis of a surgical specimen or the data of the cyto-colpo- status of the lesions, whether there is a total or partial histological monitoring obtained in at least two subsequent examinations. Among the 4,432 patients referred to our colposcopy unit between January 1985 and December 2011, because of an abnormal smear or to monitor a cervical lesion, ECCs were Correspondence to: Professor Jean-Jacques Baldauf, Department of performed by two colposcopists, exclusively in non-pregnant Obstetrics and Gynecology, Hospital of Hautepierre, 67098 patients without complete cervical stenosis. Strasbourg Cedex, France. Tel: +33 0388127472, Fax: +33 The colposcopic interpretation was made according to the criteria 0388127457, e-mail: [email protected] defined in Barcelona in 2002 by the International Federation for Cervical Pathology and Colposcopy (IFCPC) (13). Specifically, Key Words: CIN, cervical cancer, colposcopy, endocervical curettage. three situations led to ECC: i) an endocervical lesion incompletely

0250-7005/2015 $2.00+.40 4183 ANTICANCER RESEARCH 35: 4183-4190 (2015)

Table I. Correlation between diagnosis by endocervical curettage (ECC) and final endocervical diagnosis.

No. of cases ECC diagnosis

Final endocervical diagnosis Normal CIN? CIN1 CIN2 CIN3 AIS Micro SCC CAC EAC Total

Normal 208 1 14 2 225 CIN1 2 1 22 42 31 CIN2 5 2 5 30 446 CIN3 5 1 4 16 89 115 AIS 1 2 4 7 Micro 1 7 2 212 SCC 2 3 5 CAC 1 2 3 EAC 1 1 Total 222 5 45 52 106 5 2 5 2 1 445

CIN: Cervical intraepithelial neoplasia; CIN?: cervical intraepithelial neoplasia impossible to classify; AIS: adenocarcinoma in situ; Micro: microinvasive squamous cell carcinoma(SCC); CAC: cervical adenocarcinoma; EAC: endometrial adenocarcinoma. Bold values illustrate a perfect match between ECC and final diagnosis.

visible or inaccessible to directed biopsy; ii) the presence of atypical diagnosis. Patients participating in this program were informed of the glandular cells on cervical smear; and iii) a discrepancy between the registration and use of their data in accordance with the terms colposcopic appearance and the cervical smear, be it a smear with a validated and authorized by the French National Commission on Data low-grade squamous intraepithelial lesion (SIL) or atypical Processing and Liberty. The Committee for the Protection of Persons squamous cells of undetermined significance (ASC-US) without of the Strasbourg University Hospitals approved this research. T h e apparent lesion at colposcopy, or a smear with atypical squamous diagnostic reliability of ECC was compared with the final cells not excluding high-grade SIL (ASC-H), high-grade SIL, or endocervical diagnosis according to four categories: normal, mild suggestive of squamous cell carcinoma (SCC) if the colposcopic lesions (CIN1), precancerous lesions [CIN2, CIN3 or adenocarcinoma impression was not in favor of severe cervical lesion (i.e. CIN2+). in situ (AIS)], and cancer (microinvasive or invasive, squamous or We excluded six patients from this study whose tissue collected glandular). We considered ECC results as false-positives when a by endocervical curettage was insufficient for histological lesion was found, regardless of its severity, but the histological analysis, and 14 patients lost to follow-up for whom the final analysis of the surgical specimen did not confirm the presence of an histological endocervical diagnosis could not be established. endocervical lesion. Conversely, we considered as false-negative Clinical, cytological and colposcopic characteristics of the results cases in which normal tissue was found with ECC when the excluded patients were not significantly different from those of the endocervical analysis of the surgical specimen revealed a lesion. 445 patients who were included. Precancerous lesions were considered as underestimated when the All ECCs were performed using a Kervokian curette, usually ECC revealed normal tissue or CIN1. Finally, for the diagnosis of without anesthesia. To obtain sufficiently comprehensive tissue cancer, we considered as underestimated the cases in which the ECC material of the endocervical mucosa, samples were consistently showed normal tissue, CIN1-3 or AIS. When the ECC analysis taken at the four quadrants of the cervical canal and placed in a indicated CIN without being able to specify its grade, it was single collection device for histological analysis. In some very rare considered as a true-positive result for final diagnosis of CIN1, and as cases, traction on the cervix with a tenaculum was necessary to underestimated for final diagnosis of precancerous lesions or cancer. increase the accessibility of a misaligned cervix, or to introduce the curette easily in case of partial cervical stenosis. More rarely, in Statistical analysis. Descriptive statistics are tabulated by patient more severe stenosis, it was necessary to open or expand the groups. Continuous variables are summarized by means with cervical os with a scalpel or sharp scissors, which sometimes standard deviation and medians with ranges. Confidence intervals for required local anesthesia. sensitivity, specificity, positive predictive value and negative All cervical biopsies, ECCs, conization and hysterectomy predictive value of ECC were estimated by the exact binomial specimens were analyzed at the Department of Pathology of the method. Univariate analyses of the factors explaining the diagnostic University Hospital of Strasbourg. Histological findings were underestimation with ECC were made comparing continuous reported in accordance with the classification of Richart (14). variables with either the Student's t-test (when the variable of interest In accordance with the practice of our colposcopy Center, for was Gaussian), or a non-parametric test (Mann–Whitney–Wilcoxon). patients without severe cytological abnormalities, no treatment was Comparisons of qualitative variables were performed using a simple indicated in the case of a normal ECC, or in the case of CIN1 with logistic model. All tests were two-sided and a p-value of less than ECC if the colposcopy was satisfactory. In untreated patients, the 0.05 was considered significant. All variables having a level of monitoring data recorded in our Regional Organized Screening significance less than 0.1 were included in the multivariate logistic Program contributed to the validation of the final histological model. All analyses were performed with R software 3.0.2.

4184 Akladios et al: Diagnostic Value of Endocervical Curettage

Table II. Diagnostic reliability of endocervical curettage in the detection 10 cases, suggestive of SCC in 8 cases, atypical glandular of severe endocervical lesions (CIN2+) and cervical cancer. cells in 1 case, and suggestive of adenocarcinoma in 1 patient. None had a satisfactory colposcopy and the CIN2+ Cancer colposcopic aspect of the ectocervix was completely normal Prevalence, n (%) 189/445 (42.5%) 21/445 (4.7%) in 13 patients (62%). True positives, n 165 10 A directed biopsy was performed at the same time as ECC True negatives, n 248 423 in four patients with cancer (3 invasive SCCs and 1 False positives, n 8 1 microinvasive cancer stage 1). It showed an invasive SCC in False negatives, n 24 11 Sensitivity 87.3% (82.5-92.1%) 47.6% (25.8-69.4%) one patient and a CIN3 in three patients. For the 21 women Specificity 96.9% (94.7-99.0%) 99.8% (99.3-100%) with endocervical cancer, we found a perfect match in 10 Positive predictive value 95.4% (92.2-98.6%) 90.9% (73.6-100%) cases (47.6%), ECC underevaluation in 11 patients (52.4%), Negative predictive value 91.9% (87.7-94.6%) 97.5% (96.0-99.0%) with diagnosis of CIN3 in nine cases, diagnosis of AIS in one patient with adenocarcinoma, and one false-negative CIN 2+: Included CIN2, CIN3, adenocarcinoma in situ, microinvasive squamous cell carcinoma, squamous cell carcinoma, cervical result (4.8%). The latter was a 46-year-old multiparous adenocarcinoma and endometrial adenocarcinoma invading the endocervix. woman referred to colposcopy because of a high-grade SIL. ECC was performed because of a non-visible squamo- columnar junction and the absence of ectocervical or vaginal colposcopic abnormalities. The loop electrosurgical excision procedure, performed two months after the colposcopy Results because of the cyto-colposcopic discrepancy, showed an endocervical CIN3 4-mm wide associated with a The final histological diagnosis was confirmed by the microinvasion less than 1-mm deep and 1-mm wide. analysis of 255 surgical specimens (57.6%), obtained on Monitoring of 190 patients with a normal ECC or CIN1 average 2.4 months after ECC. Among the 190 women under associated with a satisfactory colposcopy, revealed a pre- surveillance only, 176 (92.6%) had normal cervical tissue cancerous lesion in four patients: one AIS diagnosed 14 with ECC and 14 (7.4%) had CIN1. The follow-up ranged months after ECC, CIN2 diagnosed in two patients 17 and from 12 to 240 months, with an average of 81.6±61.6 months 24 months after ECC, and one CIN3 diagnosed 18 months and a median of 66 months. after ECC. To determine the sensitivity, specificity and The age of the 445 patients ranged from 23 to 79 years, predictive values for ECC, we considered these lesions as with an average of 45.3 years and a median of 45 years. In underestimations or false-negative results. Table II total, 65 patients were nulliparous (14.6%), 143 were post- summarizes the reliability of ECC for the diagnosis of 189 menopausal (32.1%) and 118 patients had a history of (42.5%) severe endocervical lesions (CIN2+) and that of the cervical treatment (26.5%). Colposcopy was not satisfactory 21 endocervical carcinomas. in 367 patients (82.5%) and colposcopic impression was Among the 189 patients with severe endocervical lesions, normal in 300 patients (67.4%). Concomitant directed biopsy no clinical feature was associated with a significantly higher was performed in 113 women only (25.4%). risk of diagnostic undervaluation with ECC (Table III). The comparison between ECC diagnosis and the final Similarly, the univariate analyses showed no significantly endocervical diagnosis showed a perfect match for 362 higher risk according to the severity of the cytological patients (81.3%) (Table I). In five patients, a CIN was abnormalities, or the presence or absence of glandular cells on diagnosed with ECC but its grade could not be specified. The the referral smear. The univariate analyses showed no final diagnosis was normal cervix and CIN1 in one case, significantly higher risk of ECC underevaluation based either CIN2 in two, and CIN3 in one. Precancerous lesions were a on satisfactory or unsatisfactory colposcopy, nor on the perfect match in 123 patients (73.2%), and undervaluated colposcopic impression (Table IV). Directed biopsy revealing with ECC in 23 patients (13.5%), including false-negative severe lesion (CIN2+) was associated with a significantly results in 11 patients (6.5%). In total, 21 out of 445 ECCs lower risk of ECC underestimation [odds ratio (OR)=0.31, were performed in patients with endocervical cancer: 17 95% confidence interval (CI)=0.07-0.93; p=0.06] Conversely, SCCs including 5 of invasive cancer, 10 of stage 1 the risk of underestimation with ECC was significantly higher microinvasive cancers, 2 of stage 2 microinvasive cancer, 3 in the case of cervical cancer (OR=6.95, 95% CI=2.65-18.51; endocervical adenocarcinomas, and 1 endocervical extension p=< 0.0001). These were the only two independent factors of an endometrial adenocarcinoma. In this sub-group, 16 identified by a multivariate logistic regression analysis as patients (76%) were post-menopausal and 3 had a history of being associated with a risk of diagnostic underevaluation by conization. The cytological abnormalities on referral ECC with adjusted ORs of 0.29 (95% CI=0.06-0.92; p=0.036) Papanicolau smear were ASC-H in 1 case, high-grade SIL in and 7.33 (95% CI=2.72-20.33; p=0.00013), respectively.

4185 ANTICANCER RESEARCH 35: 4183-4190 (2015)

Table III. Univariate analyses of clinical and cytological characteristics associated with an underestimation of severe endocervical lesions with endocervical curettage (ECC).

ECC diagnosis, n (%)

Concordant or Underestimated OR (95% CI) p-Value overestimated (n = 155) (n= 34)

Age (years) mean±S D 46.8±11.6 50.0±11.1 0.14 20-39, n (%) 51 (32.9%) 7 (20.6%) 1 40-59, n (%) 82 (52.9%) 20 (58.8%) 1.78 (0.73-4.80) 0.22 60-79, n (%) 22 (14.2%) 7 (20.6%) 2.32 (0.71- 7.56) 0.16

Parity range 0-9 0-6 Nulliparous, n (%) 22 (14.2%) 5 (14.7%) 1 1, n (%) 47 (30.3%) 7 (20.6%) 0.65 (0.19-2.43) 0.51 2 and more, n (%) 86 (55.5%) 22 (64.7%) 1.13 (0.41-3.65) 0.83

Post-menopausal 57 (36.8%) 14 (41.2%) 1.20 (0.56-2.55) 0.63

Previous treatment for CIN 34 (21.9%) 4 (11.8%) 0.47 (0.13- 1.31) 0.19

Referral cytology, n (%) Glandular cells Present 130 (83.9%) 26 (76.5%) 1 Absent 25 (16.1%) 8 (23.5%) 1.60 (0.62-3.82) 0.31 Abnormalities Minor† 23 (14.8%) 6 (17.6%) 1 Major‡ 132 (85.2%) 28 (82.4%) 0.81 (0.32-2.36) 0.68

†Including low-grade squamous intraepithelial lesion (SIL) and atypical squamous cells of undetermined significance; ‡including atypical squamous cells not excluding high-grade SIL, atypical glandular cells not otherwise specified, that favor neoplasia, endocervical adenocarcinoma in situ, adenocarcinoma and squamous cell carcinoma. CIN: Cervical intraepithelial neoplasia; SD: standard deviation; OR: odds ratio; CI: confidence interval.

Discussion under-treatment of these lesions due to the significantly higher risk of underestimation of endocervical cancer with This study showed the high sensitivity of ECC for the ECC can be avoided. diagnosis of endocervical precancerous lesions and cancer. In our series, a direct diagnosis of cancer was made ECC resulted in cancer diagnosis in 96% of cases, either possible after ECC in 10 patients out of 21 (47.6%). The directly from the analysis of ECC tissue, or from that of the immediate diagnosis of invasion obtained with ECC in three cone biopsy made necessary by the ECC diagnosis of a pre- out of four patients with endocervical adenocarcinoma, and in cancerous lesion. It is worth noting that no clinical, three out of five patients with invasive SCC improved the cytological or colposcopic characteristic was associated with diagnostic and therapeutic management. In current practice, a significantly higher risk of underestimating the diagnosis this can dispense with a cone biopsy which could interfere of severe endocervical lesions with ECC, thus confirming the with the magnetic resonance imaging interpretation, especially reliability of this test in all circumstances. the accurate assessment of the size of the tumor, or complicate We recommended an excisional treatment for CINs of radical colpohysterectomy. In fact, fragile cervix or post- either high or undetermined grade identified with ECC. For conization stenosis could render the manipulation of the uterus the 167 patients with these indications, analysis of the more delicate, especially in the case of laparoscopic surgery. resected specimens confirmed severe endocervical lesion in The advantage of the immediate diagnosis of a microinvasive 157 cases (94.0%), including 10 of cancer (6.0%). These lesion is reduced since a secondary cone biopsy is absolutely results are consistent with published data indicating that necessary to confirm the absence of an associated invasive excision confirmed the presence of severe lesions (CIN2+) cancer in order to plan for adequate treatment. in three quarters of the cases, with a probability of cancer In our series, colposcopy was not satisfactory in all patients diagnosis of 2-3% (15, 16). The percentage of carcinomas with cancer. According to literature, most cases of cancer thus detected (47.6%) fully justifies our practice. Thus, discovered with ECC are associated with unsatisfactory

4186 Akladios et al: Diagnostic Value of Endocervical Curettage

Table IV. Univariate analyses of colposcopic and histological characteristics associated with an underestimation of severe endocervical lesion with endocervical curettage (ECC).

ECC diagnosis n (%)

Concordant or Underestimated OR (95% CI) p-Value overestimated (n=155) (n=34)

Unsatisfactory colposcopy 131 (84.5%) 28 (82.4%) 0.85 (0.34-2.48) 0.75

Colposcopic impression, n (%) Normal 82 (52.9%) 21 (61.8%) 1 Minor changes 15 (9.7%) 3 (8.8%) 0.78 (0.17-2.65) 0.71 Major changes 58 (37.4%) 10 (29.4%) 0.67 (0.28-1.50) 0.35

Colposcopic aspect, n (%) Normal & satisfactory 4 (2.6%) 2 (5.9%) 1 Normal & unsatisfactory 78 (50.3%) 19 (55.9%) 0.49 (0.09-3.70) 0.43 Abnormal & satisfactory 20 (12.9%) 4 (11.8%) 0.40 (0.05-3.58) 0.37 Abnormal & unsatisfactory 53 (34.2%) 9 (26.4%) 0.34 (0.06-2.70) 0.25

Directed biopsy not performed or normal, n (%) 114 (73.5%) 30 (88.3%) 1 CIN1 4 (2.6%) 1 (2.9%) 0.95 (0.05-6.72) 0.96 CIN2+ 37 (23.9%) 3 (8.8%) 0.31 (0.07-0.93) 0.06

Endocervical final diagnosis, n (%) CIN & AIS 145 (93.5%) 23 (67.6%) 1 Cancer 10 (6.5%) 11 (32.4%) 6.95 (2.65-18.51) <0.0001

CIN: Cervical intraepithelial neoplasia; AIS: adenocarcinoma in situ. CIN2+: including: CIN2, CIN3, AIS, microinvasive squamous cell carcinoma, squamous cell carcinoma, cervical adenocarcinoma and endometrial adenocarcinoma invading endocervix; OR: odds ratio; CI: confidence interval.

colposcopy (17). In some studies, the rate of cancer detected undervaluation. This was noted in 4.8% of endocervical in such patients reaches 9% (18). Many authors refer to cases carcinomas and 6.5% of pre-cancerous endocervical lesions. where only ECC helped diagnose cancer, either on immediate When the endocervical location of lesions was considered, analysis of the curetted sample or later when the cone biopsy CIN2 or 3 were observed in 8.4% of cases in a cohort of 2,857 was examined while the ECC had only shown intraepithelial women with negative ECC (1). Conversely, a false-negative neoplasia (1, 2, 5, 6, 19). In our study, only 4 out of the 21 rate of 45% corresponding to lesions diagnosed on resection patients with endocervical cancer had a directed biopsy specimens after negative curettage was reported by Andersen concomitant with ECC, indicating a high-grade CIN or a et al., for whom the rate of false-negatives was higher when cancer. This clearly shows that in the remaining 17 patients the sample was too small, usually considered as insufficient (81%), the diagnosis of cancer could be established only with by the pathologist (22). In total, among the 465 ECCs we ECC. In a study of 15 cases of cervical cancer, seven were performed, the sample was deemed inadequate or insufficient diagnosed only with ECC (20). In another study of 17 cases of for histological interpretation in only six cases, while in five cancer, six were found in patients with normal colposcopy and other cases it was not possible to specify the CIN grade. The in five cases out of these, only ECC indicated a lesion (6). Two level of sampling insufficient for histological diagnosis in our factors are associated with a higher rate of severe lesions study (2.4%) is among the lowest reported in the literature, diagnosed with ECC that would not have been detected on which range between 2 and 30% (9, 11, 15, 22-25). A rigorous directed biopsy: advanced age of the patient and unsatisfactory technique with sampling at the four quadrants of the cervical colposcopy (3, 4, 7, 12). Recently, mathematical modeling canal and a possible widening of the external os probably showed that in the case of minor cytological abnormalities, contributes to the good results of this series. There is no ECC is cost-effective only in patients over 50 years of age (21). consensus on the indications for ECC. In our colposcopy Unit, Underestimation with ECC may be due to the non-inclusion ECC concerned only 465 out of the 4,432 referred patients of the most severe part of the lesion in the tissue sample, even (10.5%), and in 332 of them without concomitant exocervical if ECC was performed at the four quadrants of the cervical biopsy. In this sub-group, due to ECC, the diagnosis of CIN2+ canal. A false-negative curettage is the most severe form of was made in 140 cases (42.2%), including 17 of cancer

4187 ANTICANCER RESEARCH 35: 4183-4190 (2015)

(5.1%). Contrary to our selective approach, almost systematic References ECC was performed during colposcopy in more than 13,000 patients, concomitant with exocervical biopsy; this resulted in 1 Moniak CW, Kutzner S, Adam E, Harden J and Kaufman RH: the additional diagnosis of CIN3+ in only 0.6% of cases, Endocervical curettage in evaluating abnormal cervical cytology. J Reprod Med 45: 285-292, 2000. including cancer in two patients (11). Similarly, performing 2 Pretorius RG, Zhang WH, Belinson JL, Huang MN, Wu LY, ECC in patients whose colposcopy was satisfactory or not, Zhang X and Qiao YL: Colposcopically directed biopsy, random Goksedef et al. diagnosed CIN2+ on ECC in 2.8% of the cervical biopsy, and endocervical curettage in the diagnosis of patients, without directed biopsy, with normal biopsy, or with cervical intraepithelial neoplasia II or worse. Am J Obstet biopsy indicating a CIN1 (19). Gynecol 191: 430-434, 2004. The follow-up of 190 patients with normal ECC or CIN1 3 Pretorius RG, Belinson JL, Azizi F, Peterson PC and Belinson with a satisfactory colposcopy during a median of 66 months S: Utility of random cervical biopsy and endocervical curettage in a low-risk population. J Low Genit Tract Dis 16: 333-338, later revealed an endocervical lesion in only four patients 2012. (2.1%), none of which was cancer. Monitoring the data 4 Solomon D, Stoler M, Jeronimo J, Khan M, Castle P and recorded within the regional campaign for organized screening Schiffman M: Diagnostic utility of endocervical curettage in ensures the completeness of the follow-up information in our women undergoing colposcopy for equivocal or low-grade series and confirms the high negative predictive value of ECC. cytologic abnormalities. Obstet Gynecol 110: 288-295, 2007. In a meta-analysis published in 1992, the negative predictive 5 Drescher CW, Peters WA, and Roberts JA: Contribution of values for ECC for microinvasive lesions were 99% for endocervical curettage in evaluating abnormal cervical cytology. satisfactory colposcopy and 96.7% in cases of unsatisfactory Obstet Gynecol 62: 343-7,1983. 6 Fine BA, Feinstein GI and Sabella V: The pre- and postoperative colposcopy (26). Other authors recommend, at least value of endocervical curettage in the detection of cervical temporarily, the follow-up of patients with minor cytological intraepithelial neoplasia and invasive cervical cancer. Gynecol abnormalities and CIN1 on ECC due to the low probability of Oncol 71: 46-49,1998. high-grade CIN and the almost certain absence of cancer (15, 7 Pretorius RG, Belinson JL, Burchette RJ, Hu S, Zhang X and 27). For these patients, the subsequent risk of CIN2+ seems Qiao YL: Regardless of skill, performing more biopsies not to be significantly higher than that observed in women increases the sensitivity of colposcopy. J Low Genit Tract Dis with CIN1 on ectocervical biopsy (28). In contrast, for CIN1 15: 180-188, 2011. 8 Driggers RW and Zahn CM: To ECC or not to ECC: the question on ECC, Gage et al. noted a significantly higher risk of remains. Obstet Gynecol Clin North Am 35: 583-597, 2008. subsequent severe neoplasia in women referred for severe 9 Goksedef BP, Api M, Kaya O, Gorgen H, Tarlaci A and Cetin cytological abnormalities compared to those referred for minor A: Diagnostic accuracy of two endocervical sampling method: cytological abnormalities (13.3% versus 3.1%) (28). randomized controlled trial. Arch Gynecol Obstet 287: 117- To our knowledge, our study based on a large series of 122, 2013. patients who underwent ECC before diagnosis of endocervical 10 Zahn CM, Rao LK, Olsen C, Whitworth SA, Washington A and pre-cancerous and cancerous lesions is the first to analyze the Crothers BA: Reproducibility of endocervical curettage clinical and cyto-colpohistological characteristics associated diagnoses. Obstet Gynecol 118: 240-248, 2011. 11 Gage JC, Duggan MA, Nation JG, Gao S and Castle PE: with ECC underestimation. Its strength comes from our Detection of cervical cancer and its precursors by endocervical standard ECC technique, and the procedure adopted for the curettage in 13,115 colposcopically guided biopsy examinations. analysis of the cone biopsy, with which the partial or total Am J Obstet Gynecol 203(481): e1-9, 2010. endocervical location of the lesion can be accurately identified. 12 Rose JD, Byun SY, Sims SM and Davis JD: The utility of The quality and duration of the follow-up provided in the endocervical curettage: does routine ECC at the time of colposcopy context of an organized screening campaign for cervical cancer for low-grade cytologic abnormalities improve diagnosis of high- guarantees the absence of overlooked lesions in patients not grade disease? Am J Obstet Gynecol 206(530): e1-3, 2012. 13 Wilkinson EJ: 2002 IFCPC World Congress. J Low Genit Tract undergoing subsequent excision. Conversely, the long Dis 6: 254- 255, 2002. colposcopic experience of the two operators raises the question 14 Richart RM: A modified terminology for cervical intraepithelial of extrapolating these good results to all colposcopists. neoplasia. Obstet Gynecol 75: 131-133,1990. 15 Petersen S, Belnap C, Larsen WI and Farley J: Grading of Conclusion squamous dysplasia in endocervical curettage specimens: The case for conservative management of mild endocervical The endocervical status of a lesion is a true difficulty for the dysplasia. J Reprod Med 52: 917-921,2007. colposcopist. In our experience, the very high sensitivity and 16 Spitzer M, Chernys AE, Shifrin A and Ryskin M: Indications for cone biopsy: pathologic correlation. Am J Obstet Gynecol 178: negative predictive value for ECC for severe endocervical 74-79,1998. lesions (pre-cancerous and cancerous) make it a very useful 17 Ferenczy A: Endocervical curettage has no place in the routine tool for reducing the number of unnecessary treatments management of women with cervical intraepithelial neoplasia: without increasing the risk of a lesion progressing to cancer. debate. Clin Obstet Gynecol 38: 644-648,1995.

4188 Akladios et al: Diagnostic Value of Endocervical Curettage

18 Krebs HB, Wheelock JB and Hurt WG: Positive endocervical 25 Ramchandani SM, Houck KL, Hernandez E and Gaughan JP: curettage in patients with satisfactory and unsatisfactory Predicting persistent/recurrent disease in the cervix after colposcopy: clinical implications. Obstet Gynecol 69: 601- excisional biopsy. MedGenMed 9: 24, 2007. 605,1987. 26 Helmerhorst TJ: Clinical significance of endocervical curettage 19 Goksedef BP, Akbayir O, Numanoglu C, Corbacioglu A, Guraslan as part of colposcopic evaluation. A review. Int J Gynecol H, Bakir LV, Kaya O, Yanik G and Cetin A: Evaluation of Cancer 2: 256-262, 1992. endocervical canal in women with minimal cervical cytological 27 Fukuchi E, Fetterman B, Poitras N, Kinney W, Lorey T and abnormalities. J Low Genit Tract Dis 17: 261-266, 2013. Littell RD: Risk of cervical precancer and cancer in women with 20 Hatch KD, Shingleton HM, Orr JW, Jr., Gore H and Soong SJ: cervical intraepithelial neoplasia grade 1 on endocervical Role of endocervical curettage in colposcopy. Obstet Gynecol curettage. J Low Genit Tract Dis 17: 255-260, 2013. 65: 403-408,1985. 28 Gage JC, Duggan MA, Nation JG, Gao S and Castle PE: 21 Shepherd JP, Guido R and Lowder JL: Should Endocervical Comparative risk of high-grade histopathology diagnosis after a Curettage Routinely Be Performed at the Time of Colposcopy? CIN 1 finding in endocervical curettage versus cervical biopsy. J A Cost-Effectiveness Analysis. J Low Genit Tract Dis 18: 101- Low Genit Tract Dis 17: 137-141, 2013. 108, 2014. 22 Andersen W, Frierson H, Barber S, Tabbarah S, Taylor P and Underwood P: Sensitivity and specificity of endocervical curettage and the endocervical brush for the evaluation of the endocervical canal. Am J Obstet Gynecol 159: 702-707, 1988. 23 Boardman LA, Meinz H, Steinhoff MM, Heber WW and Blume J: A randomized trial of the sleeved cytobrush and the endocervical curette. Obstet Gynecol 101: 426-30, 2003. 24 Klam S, Arseneau J, Mansour N, Franco E and Ferenczy A: Received April 5, 2015 Comparison of endocervical curettage and endocervical Revised April 28, 2015 brushing. Obstet Gynecol 96: 90-94, 2000. Accepted May 1, 2015

4189